Summary
The present-value cost of having a child may be at least $300K (measured in US dollars as of roughly 2012) when both direct expenditures and opportunity costs are considered. This shows the value of using the most effective birth-control methods, like the implant and vasectomy. That said, some people may find having children very important to their wellbeing, and in such cases, having children may be worth the cost.
Contents
Estimating the costs
What is the financial cost of having a child? To many people, it's the amount of money spent on food, clothing, housing, transportation, tuition, and so on. One calculator suggests that these costs might total ~$500K for a kid in the US who goes to private college, although I believe that figure does not account for the time value of money, and the estimated costs of living may be more than what frugal altruists would require. A better estimate might be something like ~$150K (assuming ~$12K of expenditures per year over 18 years at a 5% real rate of return) plus ~$60K (based on ~$160K for college ~20 years in the future using a 5% real rate of return).
Of course, economists and utilitarians will notice an additional cost: the opportunity cost of the time spent parenting. If you want to raise your child well, then this time expenditure might begin even before conception, as you research optimal health and nutrition during pregnancy. Then there are 9 months of pregnancy (which requires additional health-care visits and sickness), 
weeks or months of paternity/maternity leave from work, months of sleepless nights attending to your crying baby every two hours, talking and reading to your child daily especially during her/his first three years of life to stimulate brain developmenta, pediatrician visits, getting sick along with your kids, playing with your toddler or traveling to daycare, shopping for baby clothes and toys, researching pre / elementary / middle / high schools, helping with homework and interacting with teachers, spending quality time with your kid on the weekends, taking your kid to events and friends' houses, going to school functions, researching and visiting colleges, helping with college applications, transporting to and from college, giving advice on jobs, helping with personal-finance and legal questions, attending graduations/weddings/family gatherings, and more.
Some of this time may be well spent, either because you find it enjoyable or because you learn and grow from it. Other parts of these duties are less edifying. Say you spend 5 hours per week on chore-type work with your child,b and suppose you value what you would counterfactually have been doing in that time at $30/hour. (This doesn't mean you'd be earning $30/hour—just that the time would be worth paying that much for.) That's $150 per week, for ~19 years. Again using a 5% real rate of return for discounting, that amounts to ~$100K in present value.
There could also be costs of reduced ambition. For example, maybe you would have become a CEO who travels the world constantly, but you don't want to be the kind of mom who never sees her kid, so you aim for something less remunerative. In rare circumstances, this could add up to many millions of dollars.
And then what about your inheritance? Are you sure you're going to leave no assets to your kid, in contradiction of social custom? What if your mind changes by the time you're 70 years old?
Now, of course there might be utilitarian benefits from having a kid. The main one is that your kid might grow up to share your ideals and spend his life working on them. But I wouldn't count on it.c There is a nontrivial contribution of genes to personality and hence moral convictions, but this suggests you should become a sperm/egg donor rather than having kids yourself. And if you think environmental influence matters more, then you could go off and inspire some of the billions of other young people in the world, a minority of whom will be more receptive to your ideas than your own kid would be.
The cost of kids varies a lot by country. For example, Germany mandates 42 weeks of paid maternity leave, while the US mandates 0 weeks. It's plausible that having a child consumes more than 42 weeks of one's life, but the opportunity cost of a child is at least substantially lower in Germany, Sweden, and Norway than in the US.
Birth control
In light of the above discussion, it seems plausible that the costs of having a child could be in the ballpark of $300K. In fact, the costs could be higher depending on what you give up, but let's stick with $300K in the subsequent calculations.
With $300K at stake, birth control begins to appear pretty important. Consider the table in Wikipedia's article on "Comparison of birth control methods." Male condoms have a perfect-use failure rate of 2% per year. 2% of $300K is $6,000 per year. The combined oral-contraceptive pill has a perfect-use failure rate of 0.3% per year, or $900. And if we talk about typical-use failure rates for these methods (15% and 9%, respectively), we're into $45K and $27K per year, respectively. (That's more than some people's salaries.)
Now, these costs may be somewhat inflated, because even if pregnancy occurs, you might have an abortion, or you might leave the child for adoption. Multiply the above numbers by the probability p that you refuse to give up the child. You might think p is small, but consider that many women find abortion to be a taxing decision,d and maternal hormones promote attachment to newborns, so it's good to avoid being overconfident that you would give up the child.
Fortunately, there are some highly effective birth-control methods according to Wikipedia's table, especially "the implant," which has a typical-use failure rate of 0.05% per year, or $150*p annually. Often it's covered by health insurance.
Now, maybe there's some uncertainty in the exact failure rate, since this can vary from study to study. So instead of using 0.05% as a point estimate, suppose we take the expected value of the failure rate relative to a Bayesian probability distribution that's symmetric on the log scale. In particular, to make things really simple, suppose we think there's a 1/3 chance that 0.05% is correct, a 1/3 chance that it's 10 times too low, and a 1/3 chance that it's 10 times too high. Then the actual expected failure rate is (1/3)(0.5%) + (1/3)(0.05%) + (1/3)(0.005%) = 0.19%. This amounts to $570*p per year.
You might combine two birth-control methods together. For example, say you also do a vasectomy, with a typical-use failure rate of 0.15%. Let's again account for uncertainty and make this (1/3)(1.5%) + (1/3)(0.15%) + (1/3)(0.015%) = 0.56%. Alone, this would be $1680*p, but together with the implant, it's only $3.19*p. Note that unlike other birth-control failures, vasectomy failures are not "per year" because either the procedure succeeded indefinitely, or it failed indefinitely, with some probability (ignoring scenarios where it takes some time for the vasectomy to fail). If the vasectomy fails early, then unless you're also consistently using other birth control, you should discover that it failed as soon as a pregnancy happens.
Vasectomy
I got a vasectomy in 2012, fully paid by my health insurance. I was partly inspired by a friend who told me he had gotten the procedure. While vasectomies are not highly common in the USA, they are in other regions, such as New Zealand, where fully 25% of married men get them. The procedure was basically painless, and I had no appreciable discomfort afterwards.
In order to get approval, I had to ask my primary doctor to refer me to a urologist. My primary doctor told me I was probably making a mistake and that I might regret my decision. However, I insisted that I've known for many years that I would never want children. Ultimately he said it was my decision, so he put in the referral. Not all childless young men can get approval so easily. If your official doctor refuses, you might consider asking a local Planned Parenthood.
Will you change your mind?
It can be tricky to know whether you'll desire kids later on, and if so, how strong and long-lasting the emotion will be. It's easy to be overconfident about one's future emotions.
Robin Hanson, "Future Fertility":
Many a young woman has looked inside herself, decided that she just doesn't want kids, and went on to live her life under that assumption. But a decade or so later, her biological clock suddenly went off and she found herself very much wanting kids. I've seen this happen several times. None of us should be very confident about what introspection tells us we will later want. Evolution has designed us to express different genes at different ages; we just can't know what future genes we have been designed to express.
"Baby Fever: Does The Biological Clock Exist?":
A recent study has found baby fever is, in fact, a real thing and it affects a large number of people—men included. "Baby fever is this idea out in popular media that at some point in their lives, people get this sudden change in their desire to have children," lead researcher Gary Brase says in a release featured on The Globe & Mail. "While it is often portrayed in women, we noticed it in men, too."
On the one hand, if your desire for kids is weak and evanescent, it can be helpful to have a vasectomy or tubal ligation to prevent a temporary lapse in judgment—e.g., a baby craving that lasts a year or two but then goes away. In such cases, we might say that having kids is a permanent solution to a temporary problem. On the other hand, if you strongly crave children for many years, and are sent into despair by not being able to have them, this could be net harmful to your altruistic pursuits.
Vasectomies are often but not always reversible. Having a child is not reversible at all (unless you put the child up for adoption, though doing so is likely to make the child's life worse unless you find good parents to hand off the child to). Careful deliberation is required before either a vasectomy or having a child, but arguably, having a child is the more risky and irreversible decision. On the other hand, most people do end up having children, so one could argue that it's "riskier" to assume that you'll be in the minority that prefers not to. (As far as myself, I have an atypical personality in many ways, and I've never wanted to have children in the 33 years I've been alive as of the time I'm writing this sentence. So I'm less worried about changing my mind.)
Softening the tone
We shouldn't look down on people who decide to have kids. Many altruists will one day make that choice. Requiring childlessness as a membership criterion for the altruism club is one sure way to keep down participation rates. The important thing is just to point out the tradeoff that's being made. Having kids is not unlike, say, someone deciding to take a job that earns $120K per year instead of $140K per year. We do the best we can within the bounds of what our willpower and lifestyle sustainability allow.
The decision of whether to have children depends a lot on the individual. In "Parenthood and effective altruism," Bernadette Young describes the intense need that some people feel to have children, and the psychological difficulty that some infertile couples undergo. She encourages effective altruists to embrace the choice by some to start families rather than belittling it. While I intend this piece to highlight the costs of having children, your situation will determine whether the benefits outweigh those costs.
Bernadette's decision to have a child was highly informed and circumspect, based on knowledge of her emotions and situation. Other people have children more due to social expectations, spousal pressure, accident, or not thinking about opportunity costs. It's this latter segment of people for whom the present essay is intended.
Adoption vs. own child?
If you do have a child, should you adopt or have your own? For many people, part of the importance of having a child is having one's own child. We can see this in the fact that low-fertility couples try desperately to achieve pregnancy, even though numerous children are available for adoption. This behavior makes sense from an evolutionary perspective.
If you would be just as happy with an adopted child, the choice is less clear-cut. Adoption would probably be somewhat beneficial for the adopted child. In addition, at least in the USA, adoption is much more heavily subsidized: you can reduce your final tax bill by over $13K for adopting a child, versus $1K for your own child. On the other hand, emotional dispositions, and hence presumably moral values, are heritable to a nontrivial degree, so your own child is more likely to be altruistic than the average adopted child. Also, if you have a strong work ethic and high intelligence, your biological child is more likely to as well.
Footnotes
- Halvorson (2013-2014): "The all important issue is whether each child gets the level of stimulation—reading, talking, and direct adult interaction—that is needed to stimulate their one- and two- and three-year-old brains. [...] Most brain development and most major brain organization for each child happens in those first three years of life. [...] Reading to children in those key early years is a key tool. Speaking directly and frequently to each child and playing with each child daily in those very first years also does very many good things for children and those basic activities for each child lead the list of things that can make brains strong for babies." (back)
- The opportunity costs depend heavily on your particular situation. If your parenting time would come out of your existing leisure time, there may be no significant opportunity cost after all. If your parenting time would increase stress and require you to take even more time to mentally recuperate, the costs may be greater than what I suggested here. How much you learn (intellectually and emotionally) and what kind of person you become during the parenting process compared with the counterfactual are also relevant. (back)
- If we adopt a broad view about what types of human undertakings are socially valuable, then it becomes more plausible that a child who is smarter than average and raised in a better-than-average environment would contribute a lot more than average to future society. That said, many contributions that people make have unclear signs, especially technological progress, so it's not completely obvious if the differential talent of your child would be net good or bad on balance. (back)
- Also, given the graded nature of sentience, it's clear that a fetus also has some capacity for experiencing pain, so that even first-trimester abortions are not totally harmless, but abortion causes a small amount of suffering just compared with, say, the animals your child would eat if it were born. (back)